By Topic

Appropriate Farsi speech recognizer for commanding robots: (Performance evaluation of correlation-based and model-based classifiers for a Farsi isolated word recognition robotic system)

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Rashedi, A. ; Dept. of Electr. & Comput. Eng., Shahid Rajaee Teacher Training Univ. (SRTTU), Tehran, Iran ; Moghaddam, S.S.

In this research, two different classifier categories, correlation-based and neural network-based, are investigated for a Farsi isolated word recognizer commanding robotic system. Correlation-based category is divided to time and frequency domains. Moreover, in each of them, three decision making methods, Max, Average, and 10-Max are proposed. In addition, in neural network-based category, LPC is considered to extract the features. At first, separated samples of 4 Farsi pronounced commands (Left, Right, Forward, and Backward) go through a pre-processing section. Three methods of correlation-based category are used independently with the same data base and their performances are evaluated word by word as well as in total case. Finally the results of above mentioned methods are compared. On the other hand, LPC features get extracted independently from output of preprocessing section, and are used as inputs of the N.N. In this way one result associated to N.N.-based method is produced. Simulation results show that frequency-domain correlation-based method introduces the best recognition but, it is close to the LPC-based N.N. system. Finally it is preferred to use LPC due to lower processing time with 87.5% recognition.

Published in:

Signal Processing (ICSP), 2010 IEEE 10th International Conference on

Date of Conference:

24-28 Oct. 2010